Published in IET Radar, Sonar and Navigation Received on 22nd October 2008 Revised on 20th May 2009 doi: 10.1049/iet-rsn.2008.0171 Special Issue – selected papers from IEEE RadarCon 2008 ISSN 1751-8784 Linear and adaptive spaceborne three- dimensional SAR tomography: a comparison on real data F. Lombardini 1 M. Pardini 1 G. Fornaro 2 F. Serafino 2 L. Verrazzani 1 M. Costantini 3 1 Department of ‘Ingegneria dell’Informazione’, University of Pisa, Pisa, Italy 2 Istituto per il Rilevamento Elettromagnetico dell’Ambiente (IREA), Consiglio Naz. delle Ricerche (CNR), Napoli, Italy 3 e-GEOS S.p.A.-an ASI/Telespazio Company, Roma, Italy E-mail: f.lombardini@iet.unipi.it Abstract: Three-dimensional (3-D) synthetic aperture radar (SAR) imaging is a recent technique, based on coherent SAR data combination, and aims to obtain a full 3-D analysis in space. It is a multibaseline extension of the SAR interferometry concept and offers new options for the analysis and monitoring of ground scenes by means of the capability of separating the scattering phenomena along the height dimension. In this work, the authors summarise and extend the results obtained by processing real ERS satellite urban data characterised by a long time span of acquisition and non-uniformly spaced satellite passes, comparing the performance in height focusing obtained with a singular value decomposition (SVD)-based method and adaptive beamforming. 1 Introduction Synthetic aperture radar (SAR) two-dimensional (2D) imaging from an airborne or a spaceborne platform is a mature technique with unique capabilities in Earth observation, for example, for vegetation and snow mapping, forestry, land use monitoring, agriculture, soil-moisture determination, mineral exploration and urban mapping. A significant discrimination capability in the azimuth (along track) and range (cross track) directions is obtained by exploiting the movement of the platform carrying the radar and the coherent nature of transmitted radiation [1]. However, since the three-dimensional (3D) scene scattering properties are projected onto the 2D azimuth-range plane, possible ambiguities can arise in the inversion of physical and geometrical parameters, causing several drawbacks depending on the application. SAR interferometry is a technique that uses the phase interference of two (SAR Interferometry, InSAR) or more (multibaseline SAR interferometry, MB InSAR) views, to enable the generation of accurate digital elevation models (DEM) [1]. Unique capabilities are associated with the use of SAR MB and repeat pass data. Furthermore, the acquisition of images at different time intervals (multitemporal or multipass SAR) makes it possible the precise tracking of the velocity of ground deformations at unprecedented accuracy [2, 3]. Among other applications, InSAR has definitely given to SAR an impressive acceleration in the application to the risk management area, in particular with applications in measuring and monitoring landslides, subsidences, seismic faults and so on. Standard techniques essentially use only the phase information and assume specific models for the scattering mechanism, that is, dominated by a ‘permanent’ scatterer or distributed on the ground surface. Nonetheless, when at least one of the following conditions occurs, that is (i) the radiation penetrates under the surface (a situation that may even occur with existing sensors over specific surfaces); (ii) the ground topography is steep enough to generate critical projection of the scatterers in the slant imaging geometry (layover); (iii) the imaged area is characterised by a high 424 IET Radar Sonar Navig., 2009, Vol. 3, Iss. 4, pp. 424–436 & The Institution of Engineering and Technology 2009 doi: 10.1049/iet-rsn.2008.0171 www.ietdl.org